Thallium reinjection protocols have the disadvantage of requiring 3 sets of images, after exercise, 3-4 hr redistribution, and reinjection. An attractive alternative is to reinject 1 mCi of thallium immediately after the exercise images and imaging 3-4 hrs later, yielding a modified redistribution image representing redistribution of both the initial and the reinjected thallium doses. To determine the adequacy of exercise and modified redistribution images in identifying reversible defects, we studied 32 patients (ages 40 to 76; 23 men and 9 women) with chronic coronary artery disease by exercise thallium tomography. In addition to exercise and modified redistribution images, we also reinjected another 1 mCi of thallium after the modified redistribution image at 3-4 hrs. The exercise, modified redistribution, and 3-4 hr reinjection images were then normalized and analyzed quantitatively. Of the 81 abnormal regions on the exercise images, 30 (37%) were irreversible on modified redistribution images. However, 7 (23%) of these irreversible defects reversed after 3-4 hr reinjection, with mean thallium activities increasing from 54+5 (exercise) and 57+4 (modified redistribution) to 69+7 (reinjection, p<0.001). Only 4 of 51 reversible regions on modified redistribution appeared irreversible after 3-4 hr reinjection. All regions with reversibility only on 3-4 hr reinjection images were supplied by critically stenosed coronary arteries. These data suggest that modified redistribution imaging will overestimate myocardial fibrosis in roughly 1/4 of irreversible defects and that a 3-4 hr delay is necessary for accurate determination of myocardial ischemia and viability.